> We know GR is wrong because it is incompatible with any form of QM and we know some form of QM is more or less correct.
It’s not really about one model being correct. GR is not a perfect model because its predictions don’t match what’s observed on the scale where QM gives predictions which do.
"Wrong" is overall a poor way of thinking about models. People would like a model which is both general and elegant, not simply a model which is "right". A large and very general model with a lot of parameters which can be well tuned to fit all the observations we have would be "correct" but I am not convinced it would be very useful.
